The processes and interactions that determine the width of a river channel remain a fundamental area of investigation in geomorphology. An increasing appreciation of the capacity of riparian vegetation to alter fluvial processes, and thus influence channel form, has highlighted the need to include vegetation in these analyses. However, a disconnect exists between the small spatial and temporal scales over which the linkages among flow patterns, sediment, and plants are evaluated and the larger spatial and temporal scales in which river systems operate. In this dissertation, I strove to identify some of the key mechanisms by which vegetation affects channel width. I worked to reconcile the issue of scale by developing a novel tool that resolves patch-scale (sub-meter) patterns of hydraulic roughness over the reach scale. While the approach can be generalized to evaluate any vegetated floodplain, the multi-scalar model was specifically applied to stands dominated by the non-native riparian shrub, tamarisk, that invaded the riparian corridor of southwestern US rivers during the past century. I focused my analyses on the lower Yampa River in western Colorado. Tamarisk colonized the Yampa in the absence of other environmental perturbations. As a result, adjustments to channel form may be linked to an altered vegetation community. From a careful geomorphic and vegetation reconstruction of the Yampa, I determined that tamarisk was the driving force in channel narrowing. Application of the multi-scalar model of vegetation resistance to the Yampa enabled me to reconstruct the changing hydraulic conditions as tamarisk established and the channel narrowed over time. This hydraulic reconstruction furthered our understanding of the interactions among vegetation recruitment patterns, the increased hydraulic resistance, and the changing flow and sediment transport field. Positive feedbacks between vegetation and geomorphic change created additional areas within the channel where tamarisk could establish, and thus accelerated the rate of channel narrowing. However, these feedbacks also changed the importance of common and large floods for vegetation establishment and sediment transport. Application of this process-based understanding to future flow regimes will help managers anticipate locations along the channel that are susceptible to vegetation encroachment and changes to channel width.

I applied empirical mode decomposition (EMD) and the Hilbert-Herbert transforms, as tools to analyze streamflow data. I used the EMD method to extract and analyze periodic processes and trends in several environmental datasets including daily stream flow, daily precipitation, and daily temperature on data from the watersheds of two rivers in the Upper Colorado River Basin, the Yampa and the Upper-Green rivers. I used these data to identify forcing functions governing streamflow. Forcing functions include environmental factors such as temperature and precipitation and anthropogenic factors such as dams or diversions. The Green and Yampa Rivers have similar headwaters, but the Yampa has minimal diversions or controls while Flaming George Dam on the Green river significantly affects flow. This provides two different flow regimes with similar large watersheds. In addition to flow data, I analyzed several time series data sets, including temperature and precipitation from Northeast Utah, North Western Colorado, and Southern Wyoming. These data are from the area that defines the Yampa River and Green River watersheds, which stretch from Flaming Gorge Dam to Ouray Colorado. The EMD method is a relatively new technique that allows any time series data set, including non-linear and non-stationary datasets that are common in earth observation data, to be decomposed into a small quantity of composite finite data series, called intrinsic mode functions (IMFs). The EMD method can decompose any complicated data into several IMFs that represent independent signals in the original data. These IMFs may represent periodic forcing functions, such as environmental conditions or dam operations, or they may be artifacts of the decomposition method and not have an associated physical meaning. This study attempts to assign physical meaning to some IMFs resulting from the decomposition of the Green and Yampa flows where possible. To assign physical meaning to the IMFs, I analyzed frequencies of each IMF using the Hilbert-Hung transform, part of the Empirical Mode Decomposition method, and then compared frequencies of the IMFs with the known frequencies of physical processes. I performed these calculations on both flow, temperature, and precipitation. I found significant correlation between IMF components of flow, precipitation, and temperature data with El Niño Southern Oscillation (ENSO) events. The EMD process also extracts the long-term trend in non-linear data sets that can provide insights into the effects of climate change on the flow system. Though in preliminary stages of research, these analysis methods may lead to further understanding the availability of water within the upper Yampa and Green River Watersheds.